Linkage disequilibrium is a higher order statistical means of describing the associations of alleles at separate loci along the genome. As such it is a function of the haplotype frequencies for segments of the DNA of each chromosome. Those frequencies, in turn, reflect the accumulated influences through time from recombination, random genetic drift determined by population demographic history, selection, mutation, and chromosomal region effects. Recently, studies of a few human populations for some chromosomes and chromosome segments have suggested that linkagedisequilibrium is organized in blocks within which very high levels of association exist among alleles at all sites as a result of very few haplotypes being present. The blocks would appear to be defined by regions of low recombination bounded by highly punctate recombination. Our data from the previous funding period suggest that this picture is overly simplified with great variation around the genome and that the stochastic aspects of historical recombination and random genetic drift are very important factors. We now propose to expand those studies to cover larger segments of the genome at greater marker density in a larger number of populations from around the world. The research will involve collection of marker data on about 2500 individuals from approximately 40 populations using primarily a new SNP typing method that will allow high throughput at modest cost per marker. In total over 14 Mb of sequence will be studied, distributed as one long region (12 Mb) at 17q21 and 12 shorter regions of 200 kb each on 10 chromosomes. All regions will be covered at a density of less than 10 kb between markers. In addition to standard LD analyses, theoretical and statistical studies and computer simulations will be undertaken to improve our ability to make inferences on the factors that are primarily responsible for the patterns seen. The data will be made publicly available through ALFRED, an existing Web-accessible database.